Better questions.
Board-grade decisions.
Turn strategic ambiguity into evidence-led AI briefs, option comparisons and decision-ready executive outputs. Designed for disciplined judgment—not automated authority.
Four moves. One decision system.
Use the instruments in sequence, or enter at the part of the decision that is weakest today.
Commission the analysis
Specify the decision, success metric, evidence, output contract and accountable audience.
Make trade-offs visible
Score alternatives transparently, adjust weights and test scenario exposures.
Map the strategic thesis
Expose dependencies, assumptions, risks and the logic linking choice to execution.
Red-team before approval
Test unsupported claims, missing downside cases, decision rights and next actions.
Current decision brief
Reflects your live workspace inputs
Define the decision in Prompt Architect.
A precise decision question is the starting point for a defensible analysis.
Executive quality doctrine
Controls translated from current Claude guidance
- Explicit objectives: define what success looks like before generating.
- Grounded evidence: distinguish supplied facts, assumptions and missing data.
- Structured output: ask for options, trade-offs, recommendation and owner.
- Critical challenge: inspect alternative explanations and downside cases.
- Human approval: hold consequential actions for an authorized decision-maker.
It compiles, scores, maps and exports information locally. Copy the generated prompt into an approved Claude, ChatGPT or enterprise AI environment. Never paste unapproved confidential information into a third-party service.
Commission better intelligence.
Build a precise decision assignment with context, constraints, source discipline and an explicit board-ready output contract.
Executive mandate
Operating facts and evidence
Advanced controls · source register, examples and privacy
These are pointers supplied by you, not documents retrieved or checked by this tool.
Executive quality controls
This produces a portable prompt structure, not a fixed API configuration. For Claude, verify the current instructions for your exact model. The studio does not request hidden reasoning or rely on assistant prefills.
Assignment readiness
Structure coverage — not answer quality
Copy-ready executive prompt
Structured, reviewable, model-portable text
External tools open in a separate tab. No content is transferred automatically.
Decision laboratory.
Compare weighted strategic alternatives and three financial scenarios. Scores and probabilities are judgments entered by you—not verified predictions.
Evaluation criteria
Weights are normalized to 100% for calculation
Adjust relative importance to test strategic sensitivity.
Decision ranking
Weighted index (0–100), conditional on your inputs
Enter options and scores to compare. The highest score is not automatic approval.
Multi-criteria decision matrix
Score each option from 1 (poor) to 5 (excellent). Higher means more desirable, including risk control.
| Option | Strategic fit | Economic value | Downside protection | Speed | Execution |
|---|
You can change decision criteria weights above; scores are subjective and should be validated against underlying evidence.
Three-case financial sensitivity
Illustrative decision exposure, not an investment valuation
Probabilities must total 100%. Use a common period and consistent definition of net impact; do not mix EBITDA, revenue and NPV.
Expected outcome & governance
Updates only after all scenario inputs are valid
Strategy mind map.
Make the decision thesis, evidence, risks and execution dependencies visible. Expand, edit and export the reasoning structure.
Interactive logic map
Select any node to edit, branch or collapse it
What needs to be true?
Capture three or more assumptions that would invalidate the investment case if they failed.
What must happen first?
Trace the sequencing and ownership of decisions, funding, delivery and approvals.
What would change our mind?
Add evidence thresholds, leading indicators and plausible competitor responses.
Assurance before conviction.
Run a source-grounded challenge prompt and record management's actual review; checkmarks do not imply that AI outputs have been verified.
Executive assurance gates
Tick only after a real human check has been completed
Red-team the AI response
Generate a second-pass challenge assignment
A true independent review requires an approved AI tool or expert to assess the material. This local page generates review instructions only.
Approval record
Decision rights, reviewer identity and operating controls
This is a private working record, not an electronic signature system or a substitute for corporate approvals.
Executive playbooks.
Twelve reusable assignments that turn prompt-engineering techniques into real management deliverables.
Method translated into controls
Anthropic principles, implemented for senior executives
Primary documentation & references
External source links open only when you choose them
No model connection, live market feeds, source verification, automated assessment, encryption, authentication or cloud collaboration are claimed. Manual save uses this browser's localStorage, and JSON/TXT/SVG exports are unencrypted. For enterprise production, put identity, access policy, logging and server-side AI orchestration behind approved infrastructure.
Independent ImageFirm executive-workflow adaptation; not affiliated with or endorsed by Anthropic. Guidance reviewed 05 Nov 2026. Model-specific recommendations may change.